Genetic Alteration of the Metal/Redox Modulation of Cav3.2 T-Type

Home , Threshold potential

Genetic alteration of the metal/redox modulation of Cav3.2 T-type calcium channel reveals its role in neuronal excitability Tiphaine Voisin, Emmanuel Bourinet, Philippe Lory

To cite this version:

Tiphaine Voisin, Emmanuel Bourinet, Philippe Lory. Genetic alteration of the metal/redox modulation of Cav3.2 T-type calcium channel reveals its role in neuronal excitability. The Journal of Physiology, Wiley, 2016, 594 (13), pp.3561–74. 10.1113/JP271925. hal-01940961

HAL Id: hal-01940961 https://hal.archives-ouvertes.fr/hal-01940961 Submitted on 30 Nov 2018

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. J Physiol 594.13 (2016) pp 3561–3574 3561

Genetic alteration of the metal/redox modulation of Cav3.2 T-type calcium channel reveals its role in neuronal excitability

Tiphaine Voisin1,2,3, Emmanuel Bourinet1,2,3 and Philippe Lory1,2,3

1Centre National pour la Recherche Scientifique UMR 5203, Département de Physiologie, Institut de Génomique Fonctionnelle, UniversitédeMontpellier, Montpellier, F-34094 France 2Institut National de la SantéetdelaRechercheMédicale, U 1191, Montpellier, F-34094 France Neuroscience 3LabEx ‘Ion Channel Science and Therapeutics’, Montpellier, F-34094 France

Key points r In this study, we describe a new knock-in (KI) mouse model that allows the study of the H191-dependent regulation of T-type Cav3.2 channels. Sensitivity to zinc, nickel and ascorbate of native Cav3.2 channels is signiﬁcantly impeded in the dorsal root ganglion (DRG) neurons of this KI mouse. Importantly, we describe that this H191-dependent regulation has discrete but signiﬁcant effects on the excitability properties of D-hair (down-hair) cells, a sub-population r of DRG neurons in which Cav3.2 currents prominently regulate excitability. Overall, this study reveals that the native H191-dependent regulation of Cav3.2 channels plays a role in the excitability of Cav3.2-expressing neurons. This animal model will be valuable in addressing the potential in vivo roles of the trace metal and redox modulation of Cav3.2 T-type channels in a wide range of physiological and pathological conditions.

Abstract Cav3.2 channels are T-type voltage-gated calcium channels that play important roles in controlling neuronal excitability, particularly in dorsal root ganglion (DRG) neurons where they are involved in touch and pain signalling. Cav3.2 channels are modulated by low concentrations of metal ions (nickel, zinc) and redox agents, which involves the histidine 191 (H191) in the channel’s extracellular IS3–IS4 loop. It is hypothesized that this metal/redox modulation would contribute to the tuning of the excitability properties of DRG neurons. However, the precise role of this H191-dependent modulation of Cav3.2 channel remains unresolved. Towards this goal, we have generated a knock-in (KI) mouse carrying the mutation H191Q in the Cav3.2 protein.

The Journal of Physiology Electrophysiological studies were performed on a subpopulation of DRG neurons, the D-hair cells, which express large Cav3.2 currents. We describe an impaired sensitivity to zinc, nickel and ascorbate of the T-type current in D-hair neurons from KI mice. Analysis of the action potential and low-threshold calcium spike (LTCS) properties revealed that, contrary to that observed in WT D-hair neurons, a low concentration of zinc and nickel is unable to modulate (1) the rheobase threshold current, (2) the afterdepolarization amplitude, (3) the threshold potential necessary to trigger an LTCS or (4) the LTCS amplitude in D-hair neurons from KI mice. Together, our data demonstrate that this H191-dependent metal/redox regulation of Cav3.2 channels can tune neuronal excitability. This study validates the use of this Cav3.2-H191Q mouse model for further investigations of the physiological roles thought to rely on this Cav3.2 modulation.

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society DOI: 10.1113/JP271925 3562 T. Voisin and others J Physiol 594.13

(Received 18 November 2015; accepted after revision 29 February 2016; ﬁrst published online 2 March 2016) Corresponding author P. Lor y : In st itut de Genomique´ Fonctionnelle, 141 rue de la Cardonille, 34094 Montpellier cedex 5, France. Email: [email protected]

Abbreviations ADP, afterdepolarization potential; AP, action potential; Cav, voltage-gated calcium channel; D-hair, down hair; DRG, dorsal root ganglion; KI, knock-in; LTCS, low threshold calcium spike; LTMR, low-threshold mechanoreceptor; LVA, low-voltage activated; NT4, neurotrophin 4; TPEN, N,N,N,N-Tetrakis(2-pyridylmethyl)ethylenediamine.

Introduction (Garcia-Caballero et al. 2014) have shown antinociceptive effects in acute and/or chronic pain models. Low-voltage activated (LVA) T-type calcium channels The molecular mechanisms regulating T-type channel are critically involved in neuronal excitability, especially activity are still poorly understood. However, studies using in burst firing, in low-threshold calcium spikes (LTCS) recombinant T-type channels have established that the and rebound action potential (AP), as reviewed by Cav3.2 isoform is modulated by redox agents (Todorovic Cain & Snutch (2010). They contribute to network et al. 2001) and it is much more potently inhibited, in rhythms in normal situations, exemplified by their role in the micromolar range, by the divalent metals zinc, nickel thalamocortical circuits during sleep-related oscillations, and copper, than the Cav3.1 and Cav3.3 isoforms (Lee as well as in diseases associated with hyperexcitability et al. 1999; Jeong et al. 2003; Traboulsie et al. 2006). The disorders, such as in epilepsy and pain (Zamponi et al. histidine 191 (H191) located in the outer IS3–IS4 loop 2010; Bourinet et al. 2014). Three genes, CACNA1G, of the Cav3.2 protein is involved in the metal binding CACNA1H and CACNA1I, encode the T-type calcium site (Kang et al. 2006). This binding site also mediates channel isoforms, Cav3.1, Cav3.2 and Cav3.3, respectively the inhibition of the Cav3.2 channel by oxidizing agents (reviewed by Perez-Reyes, 2003). The corresponding and its sensitization by reducing agents, such as L-cysteine Cav3 currents display most properties of the native (Nelson et al. 2007a,b). When H191 is mutated into T-type channels, although each Cav3 isoform shows glutamine (H191Q), the channel’s affinity for zinc and some unique electrophysiological and/or pharmacological nickel is greatly reduced (Kang et al. 2006). properties. The Cav3.2 channels correspond to the In vivo studies have revealed that many of these original ‘nickel-sensitive’ LVA/T-type channels that were H191-dependent modulators of Cav3.2 channels have characterized first from the dorsal root ganglion (DRG) in marked effects on pain (Evans & Todorovic, 2015). the pioneering studies describing these channels (Carbone For example, zinc was shown to have antinociceptive & Lux, 1984; Lee et al. 1999). properties (Safieh-Garabedian et al. 1996; Liu et al. 1999; In DRGs, T-type channels play a crucial function in Nozaki et al. 2011), while reducing agents L-cysteine and tuning cell excitability. They serve as amplifiers in peri- dithiothreitol (DTT), as well as the zinc chelator N,N, pheral pain transmission in a subset of low-threshold N ,N -tetrakis(2-pyridylmethyl)ethylenediamine (TPEN) mechanoreceptor (LTMR) sensory neurons. In these display pronociceptive properties (Nelson et al. 2005, neurons, Cav3.2 channels are present in the peri- 2007b; Kawabata et al. 2007; Matsunami et al. 2011). The pheral and central terminals, as well as within the possibility that these effects are mediated through Cav3.2 axons (Talley et al. 1999; Rose et al. 2013; François channel modulation has been explored using knock-out et al. 2015; Reynders et al. 2015; Usoskin et al. 2015). animals (Choi et al. 2007; Nelson et al. 2007b). It was Cav3.2 channels in DRG neurons are involved in the tentatively hypothesized that these agents would exert patterning of afterdepolarization potentials (ADPs) and their modulatory role through their ability to directly repetitive burst firing (White et al. 1989). Activity modulate Cav3.2 channel activity and, consequently of these T-type/Cav3.2 channels is implicated in the the electrical activity of the neurons expressing these firing threshold and/or neurotransmitter release in channels. However, such demonstration clearly requires DRGs and ultimately in pain processing (Todorovic & the development and characterization of a site-specific Jevtovic-Todorovic, 2013; Bourinet et al. 2014). In a knock-in animal to investigate the physiological relevance consistent manner, knocking-down Cav3.2 expression of this H191 regulatory site in Cav3.2 channels. Toward with antisense oligonucleotides (Bourinet et al. 2005), this goal, we have created a mouse carrying the H191Q using a knock-out mouse (Choi et al. 2007), using mutation and provide here the first characterization of the selective T-type channel blockers (Todorovic et al. 2002; electrophysiological properties of the down hair (D-hair) Francois et al. 2013) or promoting Cav3.2 ubiquitination neurons of the DRG from this knock-in mouse.

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society J Physiol 594.13 H191-dependent modulation of neuronal Cav3.2 channels 3563

Methods retained and spun again at 13,000 g for 40 min at 4°C. The same amount of protein (60 μg) was loaded in Ethical approval each lane on 4–6% SDS-PAGE gels. Proteins were then The authors certify that the present work complies transferred onto nitrocellulose membranes and blocked with The Journal of Physiology’s animal checklist with 5% powdered non-fat milk. Primary antibodies (http://www.physoc.org/animal-experiments). All animal [anti-Cav3.2 (1:100, Santa Cruz Biotechnology, Inc., Santa use procedures were done in accordance with the directives Cruz, CA, USA), anti-GAPDH (1:20,000, Sigma, St Louis, of the French Ministry of Agriculture (A 34-172-41). MO, USA)] were incubated overnight in PBS-T (Tween 0.05%, milk 5%) at 4°C. After three washes in PBS-T, Generation of Cav3.2-H191Q knock-in mice secondary HRP-coupled antibodies were incubated for 1 h in PBS-T. The signal was detected using the Super Signal Mice carrying the H191Q mutation in Cav3.2 West Pico Chemiluminescent system (Pierce Chemical, were generated by the ‘Genetic engineering and Rockford, IL, USA). GAPDH was used as a loading Mouse transgenesis’ (GEMTis) facility at CIPHE control. (http://www.celphedia.eu/fr/centers/ciphe). An 11-kb genomic clone containing exon 4 of the cacna1h DRG neuron preparation gene was isolated from C57BL6/N mouse genomic DNA and cloned upon pACN vector. This clone Detailed procedures to isolate DRG neurons were was engineered to introduce the histidine-to-glutamine described elsewhere (Francois et al. 2013, 2015). Briefly, mutation (CAC to CAG) at position 191, a floxed neo adult male mice were anaesthetized with pentobarbital cassette and a thymidine kinase encoding cassette. The injection and transcardially perfused with Hank’s balanced targeting vector was linearized and electroporated into salt solution (HBSS, pH 7.4, 4°C). DRGs were dissected C57BL6/N embryonic stem cells followed by neomycin and collected in cold HBSS with 1 M Hepes, 2.5 MD-glucose selection. Selected colonies were screened for homo- and penicillin/streptomycin. DRGs were treated with a − − logous recombination by Southern blotting with Drdl mix of 2 mg ml 1 collagenase II and 5 mg ml 1 dispase digests, using 5 and 3 external probes or neo probe. for 40 min at 37°C, washed in HBSS and resuspended The floxed neo cassette contained a Cre recombinase in 1 ml of neurobasal A medium (Invitrogen, Carlsbad, under the control of a testicular promoter to achieve its CA, USA) supplemented with B27, 2 mML-glutamine auto-removal. Indeed, when chimeric males had gametes and penicillin/streptomycin. Single-cell suspensions were carrying the mutated allele in their testis, the testicular obtained by five passages through three needle tips of promoter was activated to trigger the Cre-mediated decreasing diameter (gauge 18, 21 and 26). Cells were recombination and the deletion of the floxed sequence. plated onto polyornithine/laminin-coated dishes. After ES clones with the correct genotype were micro-injected 2 h, the medium was removed and replaced with neuro- − into Balb/CN blastocysts, and the resulting chimeric basal B27 supplemented with 10 ng ml 1 neurotrophin − males were bred with C57BL6/N females to obtain 4 (NT4) and 2 ng ml 1 glial derived neurotrophic factor heterozygous F1 animals (cacna1h+/H191Q) and these F1 (GDNF). were intercrossed to obtain homozygous mutant mice (cacna1hH191Q/H191Q), named here knock-in (KI) mice. Patch clamp recordings The genotypes were determined by PCR analysis, with Patch clamp recordings were performed 6–24 h after forward primer 5-GAGAGAGGCTCAGGTTTGACA-3 plating, on neurons with a ‘rosette’ morphology that is and reverse primer 5 TGAGGCCCTCACTGGACACT-3 . promoted by the presence of NT4 in the culture media, as The 85-bp difference between KI and WT alleles resulted described by Dubreuil et al. (2004). For calcium current from the remaining loxP site in the KI allele. recordings, the extracellular solution contained (in mM): 130 choline chloride, 10 TEACl, 2 NaCl, 2 CaCl , 1 MgCl , Western blots 2 2 10 Hepes, 5 4-aminopyridine, 10 glucose (pH to 7.3 Western blots were performed as described previously with TEAOH). Pipettes with a resistance of 1–2 M (Blesneac et al. 2015). Briefly, mice were anaesthetized were filled with an internal solution containing (in mM): with sodium pentobarbital and decapitated. Whole brains 130CsCl,10Hepes,10EGTA,1MgCl2,2CaCl2, 4 Mg-ATP, were dissected, frozen in liquid nitrogen and ground with 0.3 Na-GTP (pH to 7.3 with NaOH). Whole-cell trans- a hand-held homogenizer. The samples were then placed membrane potentials were recorded in current-clamp in a lysis buffer on ice with NP-40 buffer containing 50 mM mode at room temperature. The bath solution contained Tris-HCl, pH 7.5, 125 mM NaCl, 1% NP-40, 0.1% SDS, (in mM): 140 NaCl, 4 KCl, 2 CaCl2, 2 MgCl2,10glucose,10 2mM EDTA, 2 mM EGTA and protease inhibitors (Roche Hepes (pH to 7.4 with NaOH). Pipettes with a resistance of Applied Science, Penzberg, Germany). Lysates were spun 1–2 Mwerefilled with an internalsolution containing(in at 1000 g for 10 min at 4°C. The supernatants were mM): 130 KCl, 40 Hepes, 5 MgCl2, 2 Mg-ATP,0.1 Na-GTP

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society 3564 T. Voisin and others J Physiol 594.13

(pH to 7.2 with KOH). LTCS were evoked by 2 s hyper- statistically significant at P < 0.05 (∗P < 0.05, ∗∗P < 0.01, polarizing current injections of increasing amplitude. ∗∗∗P < 0.001). The presence of the T-type current was assessed after completion of current-clamp experiments. The amplifier Results was switched back to voltage-clamp mode and the external solution was replaced by the one used for the calcium Generation of the mutant mouse carrying current recordings (see above). All drugs (purchased from the mutation H191Q on Cav3.2 Sigma-Aldrich), NiCl2,ZnCl2,ascorbate,L-cysteine and We have generated a mouse line carrying a histidine- TPEN, were applied using a gravity-driven perfusion to-glutamine point mutation at position 191 of the system. Cav3.2 subunit (Cav3.2-H191Q). As described in Fig. 1A, a C-to-G mutation was introduced using homologous Statistical analysis recombination into the cacna1h gene encoding the Cav3.2 protein of mouse (C57BL6) embryonic stem cells. The Data are presented as means ± SEMs. Statistical analyses resulting Cav3.2-H191Q knock-in mouse line (hereafter were performed using either Mann–Whitney’s test for KI mouse) was genotyped using PCR analysis with two-group comparison, or two-way ANOVA combined mutation-specific primers to reveal WT (280 bp) and with Sidak’s post hoc analysis for multiple comparisons, mutant (315 bp) alleles (Fig. 1B). Homozygous KI mice using GraphPad Prism software (GraphPad Software are viable, fertile and indistinguishable from their control Inc., La Jolla, CA, USA). P values were considered littermates. The amount of Cav3.2 protein, as determined

A B WT allele (WT) 488 bp DrdI CAC DrdI

2 3 4 5 6 7 WT/WT WT/KI KI/KI

13kb 533 bp 488 bp Modified allele H191>Q191 DrdI CAG DrdI DrdI 2 3 4* Neo 5 6 7 C

6kb 10,7kb WT KI KO α‐Cav3.2 H191Q allele (KI) 533 bp 250 kDa CAG DrdI loxP DrdI 2 3 4* 5 6 7 α‐GAPDH 55 kDa 13kb

Figure 1. Generation of a knock-in (KI) mouse line carrying the mutation H191Q on the T-type Cav3.2 channel A, the Cav3.2 allele was created by homologous recombination (see Materials and Methods). The CAC codon encoding histidine 191 (H191) in exon 4 of the WT allele (upper panel) was replaced by the codon CAG encoding glutamine (Q), and a loxP-flanked neo cassette was introduced 3 from exon 4 to select for embryonic stem cells harbouring the knock-in (KI) allele (middle panel). The final mutant allele was obtained after excision of the neo cassette by a Cre recombinase treatment of embryonic stem cells (lower panel). B, henotyping of the Cav3.2-H191Q KI mice. PCR analysis using specific primers positioned in A revealed the WT allele (488 bp) in both WT and heterozygous animals as well as the mutant allele (533 bp) in both KI and heterozygous animals. The 45 bp difference between mutant and WT alleles results from the remaining loxP site in the mutant allele. C, representative Western-blot analysis of the expression level of the Cav3.2 protein in the whole brain of WT and KI mice; a whole brain extract from knock-out mice for Cav3.2 (KO) was used as a negative control in these experiments. The expression level of α-GAPDH was used as loading control in these experiments (lower panel).

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society J Physiol 594.13 H191-dependent modulation of neuronal Cav3.2 channels 3565 by Western blot analysis, was unchanged in the whole amplitude with classical Cav3.2-like properties, including brain of KI mice compared to WT mice (Fig. 1C). half-potential (V0.5) for steady-state activation in the range of −60 mV (Fig. 2B, Table 1), V0.5 for steady-state − Altered sensitivity to zinc, nickel and ascorbate of the inactivation in the range of 72 mV (Fig. 2C, Table 1), and typical activation, inactivation and deactivation Cav3.2 current of KI mouse DRG neurons kinetics (Table 1, Fig. 2D). Importantly, no difference To determine the electrophysiological properties and could be identified regarding these T-type current electro- sensitivity of the Cav3.2 channel to its H191-dependent physiological properties recorded between D-hair neurons modulators, i.e. zinc, nickel and ascorbate, we recorded from WT and KI mice (Fig. 2, Table 1). Also, the properties whole cell T-type currents from a population of of the high voltage activated current were found to be medium-sized dissociated DRG neurons, the D-hair cells. similar (see Fig. 2B). Another hallmark property of the D-hair cells could be unambiguously identified from their T-type/Cav3.2 channel isoform is its inhibition by low typical ‘rosette-like’ morphology in short-term culture concentration of nickel (Shin et al. 2003; Dubreuil et al. conditions (see inset to Fig. 2A), allowing us to study 2004) and zinc (Traboulsie et al. 2006). To perform a a homogeneous neuron population among the various pharmacological analysis of the T-type channels in WT DRGneuronsubtypes(Dubreuilet al. 2004; Francois and KI D-hair neurons, T-type currents were recorded et al. 2013). As shown in Fig. 2A, no difference in calcium in the presence of 2 mM external calcium concentration current density could be identified between the D-hair during a depolarizing test pulse at −30 mV (90 ms neurons from WT and KI mice and this current was duration) from a holding potential at −80 mV (see efficiently blocked (>90%) in both conditions in the Fig. 3A). Figure 3B shows the dose–response of zinc presence of 10 μM TTA-A2, a selective T-type channel inhibition of Cav3.2 current in D-hair neurons from WT blocker (Francois et al. 2013). D-hair cells are mechano- mice and from KI mice. The IC50 for zinc inhibition of receptor neurons that express Cav3.2 current of large the WT channel was 1.7 ± 0.4 μM (n = 7), contrasting

A B )

–1 50 V 0.2 m (mV) 40 –80 –60 –40 –20 20 40

30 WT (13) –0.2 KI (12) 20 –0.4

Figure 2. Electrophysiological Current density pF (pA properties of T-type currents in DRG –0.6 neurons from WT and KI mice 10 A, T-type calcium current density in D-hair (41) (46) (6) (7) –0.8 DRG neurons from wild-type (WT) and

0 max I

knock-in (KI) mice, in control conditions / WT KI WT KI –1.0 I (left) and in the presence of 10 μM of the T-type channel inhibitor TTA-A2 (right). For Control TTA‐A2 10 μM each cell, the current amplitude (in pA) was measured for a depolarization step from C D −80 to −30 mV and normalized for the cell capacitance (pF) in current density 25 (pA pF–1). Values are expressed as 1.0 WT (8) WT (8)

max KI (8) KI (8) ± I mean SEM. Inset: bright ﬁeld image of a / 20 I typical D-hair neuron with its ‘rosette’ morphology. Scale bar = 30 μM. B, 15

normalized I–V relationship for the WT deactivation (ms) τ (grey square) and the KI (black circle) 0.5 current evoked by 90 ms depolarizing step 10 pulses from a holding potential at −80 mV. C, steady-state inactivation curves obtained 5 by stepping the membrane potential at −30 mV from conditioning depolarizing pulses ranging from −100 to −30 mV. D, 0 0 –100 –90 –80 –70 –60 –50 –40 –30 –120 –100 –80 –60 deactivation kinetics as a function of the voltage. Potential (mV) Potential (mV)

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society 3566 T. Voisin and others J Physiol 594.13

Table 1. Electrophysiological properties of T-type currents in performed current-clamp recordings on freshly cultured D-hair neurons from WT and KI mice D-hair neurons. Action potentials (APs) were evoked by 4 ms current injections of increasing intensity. Typical WT KI recordings of evoked APs are presented in Fig. 4A for a Activation WT neuron (upper panel) and for a KI neuron (lower V0.5 (mV) −58.1 ± 0.77 (13) −60.27 ± 0.73 (12) panel), in control condition (left part) and in the presence Slope (mV) 2.68 ± 0.21 (13) 2.3 ± 0.27 (12) of 30 μM nickel (right part). Under control conditions, the Inactivation rheobase (threshold current) was similar in WT neurons − ± − ± V0.5 (mV) 72 1.02 (8) 71.06 1.05 (8) (0.82 ± 0.06 nA, n = 15) and KI neurons (0.71 ± 0.06 nA, ± ± Slope (mV) 4.88 0.25 (8) 4.5 0.3 (8) n = 13). After application of 30 μM nickel, the rheobase Activation kinetics was significantly increased in WT neurons (0.9 ± 0.06 nA, τ − ± ± @ 50 mV (ms) 8.44 0.43 (13) 7.2 0.56 (12) n = 15), consistent with previously described findings Inactivation kinetics τ @ −50 mV (ms) 17.73 ± 0.62 (13) 16.8 ± 0.53 (12) (Dubreuil et al. 2004), whereas it remained unchanged in KI neurons (0.72 ± 0.05 nA, n = 13) (Fig. 4B). Of Data are presented as mean ± SEM with (n) the number of cells interest, the AP in mouse D-hair neurons is characterized tested. Note that none of the values was statistically different by a spike followed by a pronounced ADP and this ADP between WT and KI conditions. is dependent on the activity of T-type Cav3.2 channels (Figs4and5;seealsoDubreuilet al. 2004). Noticeably, with the IC50 determined for the KI channel which was the ADP amplitude was significantly larger in KI neurons ± μ = 10 2 M (n 5). We also assessed the effect of other as compared to WT neurons (Fig. 5B). Application of zinc H191-dependent modulators on neuronal Cav3.2 current (Zn; Fig. 5C) and nickel (Ni; Fig. 5D)atlowconcentration from the KI mouse. Application of the oxidizing agent (1 μM) was sufficient to induce a strong inhibition of μ ± = ascorbate at 300 M blocked 27.9 1.9% (n 7) of the ADP amplitude on WT neurons (65 ± 35%, n = 6; ± = WT current and only 8.1 1.5% (n 9) of KI current 77 ± 25%, n = 7, respectively). Importantly, inhibition of (Fig. 3C). Also, the blocking effect of nickel ions was the ADP amplitude in the presence of 1 μM zinc or nickel significantly reduced for the Cav3.2 current recorded on was lost on KI neurons. A higher concentration of zinc ± D-hair neurons from KI mice. Nickel blocked 29.5 2.7% (30 μM) completely abolished the ADP in WT neurons and μ ± μ of WT current at 1 M and 79.8 0.6% at 30 M,whereas also significantly reduced ADP amplitude in KI neurons ± ± it blocked respectively 12.9 1.8 and 28 2.3% of (Fig. 5C). In contrast, 30 μM nickel abolishes the ADP in = KI current (n 8) (Fig. 3C). In good agreement with WT neurons but had no effect on KI neurons (Fig. 5D). the absence of free zinc in our standard experimental ThesedatashowthatADPpropertiescanbemodulatedby μ conditions, neither the zinc chelator TPEN (10 M)nor H191-dependent regulation of T-type Cav3.2 channels. the reducing agent L-cysteine (100 μM) could modulate the T-type current density in both WT neurons (93.5 ± 1.5%, Properties of the LTCS in DRG neurons of KI mice n = 7, and 91.9 ± 1.2%, n = 6, for TPEN and L-cysteine, respectively) and KI neurons (95.4 ± 1.8 %, n = 8, and To further examine the neuronal excitability in the D-hair 96.0 ± 2.0%, n = 7, for TPEN and L-cysteine, respectively; neurons from the KI mice, we investigated their ability Fig. 3D). Importantly, similar experiments performed in to generate LTCS. It was previously shown that LTCS the presence of 1 μM zinc revealed the ability of both properties rely on T-type channel properties (Dubreuil TPEN and L-cysteine to unmask a basal inhibition of WT et al. 2004; Francois et al. 2013). Using 2 s hyper- Cav3.2 channels by free extracellular zinc (121.5 ± 2.1%, polarizing steps of increasing amplitude, in the range −70 n = 7, for TPEN and 122.4 ± 2,3%, n = 6, for L-cysteine to −100mV,led totheappearanceoftypicalLTCScrowned in WT neurons and 104 ± 1.4%, n = 7, for TPEN with a single sodium channel-dependent AP (Fig. 6A). and 97.9 ± 2.4%, n = 6, for L-cysteine in KI neurons; In control conditions, the membrane potential necessary Fig. 3E). Collectively, these data demonstrate that the to trigger an AP was not statistically different between Cav3.2 channels present in neurons from mice carrying the WT and KI neurons (WT: −88.2 ± 1.8 mV, n = 7; H191Q mutation exhibit a reduced sensitivity to the main KI: −94.2 ± 2.0 mV, n = 9, P = 0.09). Application representative pharmacological agents known to modulate of 1 μM zinc prevented the triggering of an AP for the Cav3.2 channels through the H191 residue. same hyperpolarization level in WT neurons (Fig. 6A, upper left panel, Fig. 6B). This blocking effect on the Properties of the action potentials in DRG neurons threshold of LTCS was lost in KI neurons (Fig. 6A, from KI mice lower left panel, Fig. 6B). Regarding LTCS amplitude, which was measured in the range 5–10 ms after the To evaluate the role of H191-dependent regulation AP peak as the difference from the resting membrane of the Cav3.2 channels on neuronal excitability, we potential, it was unchanged between WT neurons and

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society J Physiol 594.13 H191-dependent modulation of neuronal Cav3.2 channels 3567

KI neurons in control conditions (Fig. 7A, B). However, from the KI mouse appear more excitable due to a loss LTCS amplitude was strongly inhibited following the of inhibition, which indicates that the H191-dependent application of 1 μM zinc in WT neurons but only regulation of Cav3.2 channels can influence the excitability moderatelyreducedinKIneurons.TheLTCSinhibitionin of neurons expressing these channels. WT neurons was significantly higher than in KI neurons (Fig. 7C). Similar experiments were conducted in the pre- Discussion sence of nickel (Fig. 7D). LTCS amplitude in WT neurons was significantly reduced by increasing concentrations of In this study, we describe a Cav3.2-H191Q knock-in (KI) nickel (1 and 30 μM), in contrast to LTCS amplitude in mouse model that allows the study of the H191-dependent KI neurons which was unchanged following application regulation of Cav3.2 channels. The sensitivity to zinc, 30 μM nickel. Together our data show that DRG neurons nickel and ascorbate of native Cav3.2 channels is

A B 1.0 WT *** 0.8 KI ( ) IC = 10.0 μM + Zn 10 μM 1nA 50 + Zn 1 μM 10 ms 0.6 *** Control max I / I 0.4 () KI WT *** IC50= 1.7 μM 0.2

+ Zn 10 μM + Zn 1 μM 0 Control –8 –6 –4 –2 [Zn] log (M) CDE

140 WT 140 140 KI ** *** (%) (%)

120 120 ) (%) 120 Ctrl Ctrl M I I / / I I 100 *** *** *** 100 100

80 80 80 (@zinc1μ Ctrl

60 60 I 60 / I 40 40 40

20 20 20 (7) (9)(8) (8) (8) (8) (7)(8) (6) (7) (7)(7) (6) (6) 0 0 0 Asc. Ni Ni TPEN L‐cys. TPEN L‐cys. 300 μM 1 μM 30 μM 10 μM 100 μM 10 μM 100 μM

Figure 3. Metal/redox sensitivity of T-type currents in DRG neurons from WT and KI mice A, representative T-type current traces recorded from D-hair cells isolated from a WT mouse (upper trace) and a KI mouse (lower trace) evoked by 90 ms depolarization steps from −80 to −30 mV, before and after exposure to 1 and 10 μM zinc. B, dose–response curves obtained for zinc inhibition of T-type currents in neurons from WT mice and KI mice. Current amplitude was normalized to the peak current in the absence of zinc, and the remaining current was plotted against zinc concentration. The IC50 values were obtained from ﬁtted data using a sigmoidal dose–response with variable Hill slope equation. Values are mean ± SEM (∗∗∗P < 0.001; two-way ANOVA). C, the ability of ascorbate (300 μM) and nickel ions (1 and 30 μM) to inhibit T-type currents in D-hair neurons from WT (black bars) and KI (grey bars) mice is represented as the percentage of remaining current following drug application. D, effect of zinc chelator TPEN (10 μM)andL-cysteine (100 μM) on T-type current. Note that TPEN and L-cysteine in control conditions fail to modulate either the WT or the KI current. WT TPEN vs. KI TPEN: P = 0.52; WT L-cysteine vs. KI L-cysteine: P = 0.0734; Mann–Whitney). E, effect of TPEN (10 μM)andL-cysteine (100 μM) on the T-type current from WT and KI mice in the presence of 1 μM zinc. Values are expressed as mean ± SEM (∗∗P < 0.01; ∗∗∗P < 0.001; Mann–Whitney).

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society 3568 T. Voisin and others J Physiol 594.13 significantly impeded in the DRG neurons of this KI metal-binding site at the external surface of the Cav3.2 mouse. We report that this H191-dependent regulation channel. This metal-binding site comprises not only the has discrete but significant effects on the excitability H191 residue in the extracellular IS3–IS4 loop (Kang properties of D-hair cells, a subpopulation of DRG et al. 2006) but also D189 and G190 in the IS3–IS4 neurons in which Cav3.2 currents play a prominent role loop as well as D140 in the outer IS2 segment (Kang in regulating excitability. et al. 2010). Although the protein sequence alignment Because there was previously no means to assess directly indicates that these four residues are conserved among the relevance of the H191-dependent modulation of native most species, including mouse and human (Uniprot Cav3.2 channels, as reviewed by Evans & Todorovic Nos. AAK21607 and NP 001005407, respectively, not (2015), we have generated a mouse model harbouring shown), one neighbouring neutral residue, alanine (A141) an H191Q point mutation in its cacna1h gene. In this in the human sequence, is substituted by a negatively Cav3.2-H191Q KI mouse model, we found no change charged residue, aspartate (D141), in the mouse Cav3.2 in the Cav3.2 protein expression level, in the Cav3.2 protein. Of note, when expressed in Xenopus oocytes, current density and in the electrophysiological properties. the A-to-D mutation introduced into the human Cav3.2 Substitution of the H191 residue in the Cav3.2 iso- protein (A141D) was found to increase the IC50 value form by glutamine (Q), the corresponding residue in the for zinc inhibition from 3 to 12.5 μM (Kang et al. nickel/zinc-resistant Cav3.1 isoform, has proven efficient 2010). Also, post-translational maturation of the Cav3.2 in significantly reducing the nickel/zinc sensitivity of channel in neurons and/or association with putative the Cav3.2 isoform in heterologous expression studies Cav3.2-interacting proteins might be different in neurons (Kang et al. 2006). Of importance, the H191Q mutation and in HEK-293 cells, providing a distinct environment also disrupts the metal/redox-dependent regulation of of the metal-binding site among mammalian cell types Cav3.2 channels in KI mice. However, compared to studies able to modify the zinc/nickel affinity for the channel. performed with recombinant Cav3.2 channels (human Nevertheless, our study establishes that in the range isoform) expressed in HEK-293 cells (Traboulsie et al. 1–30 μM of zinc and nickel, there is a significant difference 2006; Nelson et al. 2007b), we found that the IC50 for in current inhibition between neuronal WT and H191Q zinc inhibition was 2-fold higher in WT mouse neurons Cav3.2 channels which therefore validates that the H191 (1.7 μM) and 4-fold lower in KI mouse neurons (10.0 μM), residue in Cav3.2 channels is an important component of compared to that obtained with human recombinant WT the zinc/nickel binding site, as reported with recombinant Cav3.2 and Cav3.2-H191Q channels (0.8 and 42 μM, Cav3.2 channels. respectively). This difference in zinc sensitivity between D-hair neurons are a population of mechanoreceptor the native and recombinant Cav3.2 channels could be DRG neurons that express a large density of Cav3.2 current related to a difference in the primary sequence of the (Shin et al. 2003; Dubreuil et al. 2004; Wang & Lewin, mouse Cav3.2 protein, which is not identical to its human 2011; Hilaire et al. 2012). These channels represent a counterpart. Indeed, zinc binds to a complex high affinity main route for calcium entry and are associated with

A B ns WT 1.5 * ns

20 mV 4 ms Figure 4. Action potential (AP) properties in DRG neurons from WT and KI mice Rheobase (nA) 1.0 A, typical traces of evoked APs. + Ni 30 μM Depolarization currents of increasing amplitude were injected to trigger an AP in the absence (left) or presence of 1 μM zinc (right) in D-hair neurons from WT (upper KI panels) and KI (lower panels) mice. The 0.5 resting potential of these neurons was maintained at −70 mV. B, threshold current for an AP generation (Rheobase) in control condition and with 30 μM nickel for neurons + Ni 30 μM from WT mice (left) and KI mice (right). 0.0 Statistical analysis was performed with Ctrl Ni Ctrl Ni repeated-measures two-way ANOVA (WT Ctrl μ ∗ < 30 μM 30 μM vs. WT Ni 30 M, P 0.05; WT Ctrl vs. KI Ctrl, P = 0.52; WT Ni 30 μM vs. KI Ni 30 μM, WT KI P = 0.08; KI Ctrl vs. KI Ni 30 μM, P = 0.95).

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society J Physiol 594.13 H191-dependent modulation of neuronal Cav3.2 channels 3569

AP generation and electrical activity. In Cav3.2 knock-out observed in WT neurons. A similar differential effect was mice, D-hair neurons showed lower firing frequency and found following application of 1 μM zinc, although the reduced mechanosensitivity, compared to WT mice (Wang ADP was markedly inhibited using a higher concentration & Lewin, 2011). Establishing that the T-type current in (30 μM). The contribution of H191-dependent regulation D-hair cells was significantly less sensitive to zinc and of Cav3.2 channels to the LTCS was also identified: (1) nickel in the KI mouse allowed us to probe the role of in the presence of low zinc or nickel concentration, the H191-dependent modulation of the Cav3.2 channel in the hyperpolarization potential necessary to trigger an shaping the AP of D-hair neurons. Using current-clamp LTCS crowned by a single spike as well as the LTCS experiments combined with nickel/zinc pharmacological amplitude were unchanged in KI D-hair neurons, while analysis we found that the rheobase threshold current was significantly increased and reduced, respectively, in WT significantly increased in the presence of 30 μM nickel in neurons. Together, these data confirmed previous studies WT D-hair neurons, an effect that was lost in KI D-hair on the important role of Cav3.2 channels on the AP neurons. A particularity of the AP waveform in D-hair and LTCS properties of D-hair neurons and demonstrate neurons is the presence of ADP that is dependent on the that the H191-dependent regulation of these channels can T-type channels (Dubreuil et al. 2004). Here we describe contribute to the D-hair neuron excitability. that in the presence of 1–30 μM nickel, the ADP was In our current-clamp experiments, we have identified a unchanged in the KI D-hair neurons, contrary to that differential ability of nickel and zinc ions to modulate

A B 25 WT KI * 20 20 mV ADP (mV) 15 10 ms 10 Control Control Zn 1 μM Zn 1 μM 5

Zn 30 μM (18) (21) Zn 30 μM 0 WT KI C D

25 25 WT *** KI 20 20 *** *** 15 * 15

10 10

5 5 ADP (mV) ADP (mV)

0 0

–5 –5 (6)(6) (6) (7) (7) (7) (12)(7) (12) (13) (7) (13) –10 –10 Ctrl Zn Zn Ctrl Zn Zn Ctrl Ni Ni Ctrl Ni Ni 1μM 30 μM 1μM 30 μM 1μM 30 μM 1μM 30 μM

Figure 5. Afterdepolarization (ADP) properties in DRG neurons from WT and KI mice A, representative APs with ADP before and after application of zinc (1 and 30 μM) in D-hair neurons of WT (left) and KI mice (right). The resting membrane potential of the neurons was maintained at −70 mV. B, ADP amplitude (in mV) in control condition is determined as the difference in voltage from the resting potential value ∗ ( P < 0.05; Mann-–Whitney). C and D, effect of the application of zinc at 1 and 30 μM (C) and nickel at 30 μM (D) on ADP amplitude. Statistical analysis was performed with two-way ANOVA followed by Sidak’s test (∗P < 0.05; ∗∗∗ P < 0.001; KI Ctrl vs. KI Zn 1 μM, P = 0.22; WT Ctrl vs. WT Ni 1 μM, P = 0.11; KI Ctrl vs. KI Ni 1 μM, P = 0.61; KI Ctrl vs. KI Ni 30 μM, P = 0.43).

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society 3570 T. Voisin and others J Physiol 594.13

D-hair cell excitability. We found that AP and LTCS Glra1-D80A neurons as compared to WT neurons. of D-hair KI neurons were not, or poorly, affected by This study also provides evidence that the Glra1-D80A low concentrations of nickel but partly altered with knock-in mice exhibit severe motor deficits and a hyper- corresponding concentrations of zinc. Indeed, zinc is an ekplexia phenotype mimicking the hereditary Startle important trace metal that is involved in many biochemical disease. Using acute hippocampus and spinal cord slices and physiological processes in mammalian tissues (Mathie from NR2A-H128S mice, Nozaki et al. (2011) reported et al. 2006) and free zinc modulates many membrane that NMDA excitatory postsynaptic currents (EPSCs) were receptors and channels, such as NMDA receptors (Paoletti barely affected by a submicromolar zinc concentration, et al. 1997; Nozaki et al. 2011), glycine receptors (Nevin contrary to that observed with WT mouse slices in which et al. 2003), TRPA1 (Hu et al. 2009), ASIC channels NMDA EPSCs were markedly inhibited. They also found (Baron et al. 2001), potassium channels (Easaw et al. that, using pain models of inflammatory pain (complete 1999; Zhang et al. 2001; Clarke et al. 2004; Prost et al. Freund’s adjuvant) and neuropathic pain (sciatic nerve 2004), sodium channels (Sheets & Hanck, 1992; Kuo ligation), NR2A-H128S animals developed enhanced et al. 2004) and other calcium channels (Magistretti et al. mechanical allodynia with loss of zinc-induced analgesia. 2003) especially Cav2.3 (Shcheglovitov et al. 2012). Taken together, these landmark studies emphasize how There is a need for such a site-specific knock-in important is the development of dedicated animal mouse model to investigate the physiological relevance models to unravel the modulation properties and physio- of H191-dependent regulation of Cav3.2 channels and we logical relevance of trace metal regulation of channels believe that this new animal model will be instrumental and receptors and how it critically impacts neuronal to explore the role of this modulation in physiological excitability. situations. Indeed, a few other site-specific knock-in A current challenge is to identify the ability of end- animal models dedicated to the study of the physio- ogenous trace metals, especially zinc, to modulate tonically logical relevance of zinc modulation have been developed and/or dynamically native channels and receptors. Using recently, including (1) a Glra1-D80A knock-in mouse line hippocampal slices from NR2A-H128S animals, it was lacking a zinc binding site in the glycine receptor (GlyR) α1 established recently that an ambient zinc level at mossy subunit gene (Hirzel et al. 2006) and (2) an NR2A-H128S fibre–CA3 synapses was not high enough to tonically knock-in mouse line lacking the high-affinity zinc-binding occupy a high-affinity zinc site on NMDA receptors in site in the NR2A subunit of NMDA receptor (Nozaki the presence of physiological extracellular magnesium et al. 2011). Using Glra1-D80A animals, Hirzel et al. concentrations (Vergnano et al. 2014). Regarding Cav3.2 (2006) found a significant reduction in the potentiation channels, evidence for tonic inhibition by zinc was of glycine currents by low concentrations of zinc in found in peripheral C-type nociceptors that also express

A B ns

** ns –80 WT 20 mV 20 ms –85 Figure 6. Rebound activity, low + Zn 1 μM threshold calcium spike (LTCS) and its modulation by zinc in DRG neurons from WT and KI mice –90 A, typical traces of the rebound potential LTCS evoked following 2 s hyperpolarizing steps of increasing amplitude in D-hair KI –95 neurons in control (left panel) and in the presence of zinc (1 μM, right panel) from WT

LTCS thresholdLTCS (mV) (upper panel) and KI (lower panel) mice. B, LTCS threshold in control condition (circle) –100 and in the presence of zinc (1 μM,square)in + Zn 1 μM D-hair neurons from WT mice (black symbols) and KI (grey symbols) mice. Statistical analysis was performed with –105 repeated-measures two-way ANOVA Ctrl Zn Ctrl Zn ∗∗ ( P < 0.01; KI Ctrl vs. KI Zn 1 μM, P = 0.99; 1 μM 1 μM WT Ctrl vs. KI Ctrl, P = 0.09; WT Zn 1 μM vs. WT KI KI Zn 1 μM, P = 0.70).

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society J Physiol 594.13 H191-dependent modulation of neuronal Cav3.2 channels 3571 large Cav3.2 T-type currents (Nelson et al. 2007b). In sulfide (H2S) and carbon monoxide (CO), could also their study, Nelson et al. (2007b) showed that reducing play a role in Cav3.2 channel regulation through agents, such as L-cysteine, as well as synthetic and end- redox-related mechanisms involving H191 (Boycott et al. ogenous zinc chelators, diethylenetriaminepenta-acetic 2013; Elies et al. 2014; Sekiguchi et al. 2014). This redox acid (DTPA), TPEN, DTT and BSA, could sensitize Cav3.2 modulation of Cav3.2 channels could also be involved current and favour AP firing in C-type nociceptors from in the GABAB receptor inhibition of T-type channels WT mice but not from Cav3.2 knock-out mice. These in DRG neurons (Huang et al. 2015). Of interest, it experiments suggested that in WT animals, zinc chelation was recently reported that chelating zinc with TPEN could effectively remove Cav3.2 channel tonic inhibition, could facilitate modulation of GABAergic inhibitory favouring thermal hyperalgesia (Evans & Todorovic, post-synaptic currents in hippocampal granule cells, an 2015). However, because most reducing agents used effectthatwasoccludedwhenblockingT-typechannels have non-specific chelating abilities, channel modification (Grauert et al. 2014). Increased Cav3.2 channel activity due could occur by chelation of trace metals or through to the removal of a chelatable zinc fraction would enhance alternative mechanisms, such as reduction of cysteine cellular excitability in fast-spiking interneurons without residues. In addition, other studies have suggested that affecting the properties of granule cells, supporting that the endogenous gasotransmitters, such as hydrogen zinc modulation of GABAergic transmission to granule

A B 25 ns

20 WT KI 20 mV 15 20 ms Control Control 10 + Zn 1 μM + Zn 1 μM 5 LTCS amplitude (mV) LTCS (15) (16) 0 WT KI C D *** *** *** *** * 1.5 *** 1.5 ** WT KI 1.0 1.0

0.5 0.5

0.0 0.0

–0.5 –0.5 LTCS amplitude (normalized) LTCS LTCS amplitude (normalized) LTCS (13) (8) (5) (11) (10) (8) (7) (7) (7) (6) (6) (6) –1.0 –1.0 Ctrl Zn Zn Ctrl Zn Zn Ctrl Ni Ni Ctrl Ni Ni 1 μM 30 μM 1 μM 30 μM 1 μM 30 μM 1 μM 30 μM

Figure 7. Zinc modulation of LTCS amplitude in DRG neurons from WT and KI mice A, typical traces of LTCS evoked by a 2 s hyperpolarizing pulse in D-hair neurons from WT (left) and KI mice (right) before and after exposure to 1 μM zinc. B, averaged amplitude of the LTCS in DRG neurons from WT and KI mice. The amplitude of the LTCS is measured as the difference from the resting membrane potential value (P = 0.27; Mann–Whitney). C, normalized amplitude of the LTCS before and after exposure to 1 and 30 μM zinc. Statistical analysis was performed with two-way ANOVA followed by Sidak’s test (∗∗∗P < 0.001; KI Ctrl vs. KI Zn 1 μM, P = 0.19). D, normalized amplitude of the LTCS before and after exposure to 1 and 30 μM nickel. Statistical analysis was performed with two-way ANOVA followed by Sidak’s test (∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; KI Ctrl vs. KI Ni 1 μM, P = 0.96).

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society 3572 T. Voisin and others J Physiol 594.13 cells would involve H191-dependent regulation of Cav3.2 Cain SM & Snutch TP (2010). Contributions of T-type calcium channels. channel isoforms to neuronal firing. Channels 4, 44–51. The major finding of this study is that impairment of Carbone E, Calorio C & Vandael DH (2014). T-type the native H191-dependent regulation of T-type Cav3.2 channel-mediated neurotransmitter release. Pflugers Arch channels reveals it potential role in the excitability of 466, 677–687. Carbone E & Lux HD (1984). A low voltage-activated, fully Cav3.2-expressing neurons. This study opens new avenues inactivating Ca channel in vertebrate sensory neurones. for precise investigation of the role of the H191-dependent Nature 310, 501–502. regulation of T-type Cav3.2 channels. Over the last decade, Choi S, Na HS, Kim J, Lee J, Lee S, Kim D, Park J, Chen CC, a role of Cav3.2 channels in a variety of physiological Campbell KP & Shin HS (2007). Attenuated pain responses functions has been proposed (Huc et al. 2009). This in mice lacking CaV3.2 T-type channels. Genes Brain Behav includes pain detection and processing (Todorovic & 6, 425–431. Jevtovic-Todorovic, 2007; Bourinet et al. 2014), neuronal ClarkeCE,VealeEL,GreenPJ,MeadowsHJ&MathieA excitability and pacemaker activities (Liao et al. 2011), (2004). Selective block of the human 2-P domain potassium hormone secretion (Maturana et al. 2009; Enyeart & channel, TASK-3, and the native leak potassium current, Enyeart, 2015), neurotransmitter release (Weiss et al. 2012; IKSO, by zinc. JPhysiol560, 51–62. Carbone et al. 2014), vascular tone (Hansen, 2015) and Dubreuil AS, Boukhaddaoui H, Desmadryl G, Martinez-Salgado C, Moshourab R, Lewin GR, Carroll P, fertilization (Bernhardt et al. 2015). The Cav3.2-H191Q Valmier J & Scamps F (2004). Role of T-type calcium current mouse model will be helpful to decipher a putative role in identified D-hair mechanoreceptor neurons studied in of this Cav3.2 modulation, not only in the physiological vitro. JNeurosci24, 8480–8484. processes listed above, but also in a variety of disease Easaw JC, Jassar BS, Harris KH & Jhamandas JH (1999). Zinc states that are thought to rely on altered Cav3.2 activity, modulation of ionic currents in the horizontal limb of the such as neuropathic pain, epilepsy, primary aldosteronism diagonal band of Broca. Neuroscience 94, 785–795. and cardiac hypertrophy (Evans & Todorovic, 2015). Elies J, Scragg JL, Huang S, Dallas ML, Huang D, MacDougall D, Boyle JP, Gamper N & Peers C (2014). Hydrogen sulfide Future studies should benefit from the use of these + transgenic animals lacking H191 to address in vivo the inhibits Cav3.2 T-type Ca2 channels. FASEB J 28, potential roles of the trace metal and redox modulation 5376–5387. of T-type channels in a wide range of physiological Enyeart JJ & Enyeart JA (2015). Adrenal fasciculata cells express T-type and rapidly and slowly activating L-type Ca2+ and pathological conditions, as well as to support the channels that regulate cortisol secretion. Am J Physiol Cell development of selective Cav3.2 channel modulators Physiol 308, C899–918. targeting the H191-dependent regulation. Evans JG & Todorovic SM (2015). Redox and trace metal regulation of ion channels in the pain pathway. Biochem J References 470, 275–280. Francois A, Kerckhove N, Meleine M, Alloui A, Barrere C, Baron A, Schaefer L, Lingueglia E, Champigny G & Lazdunski Gelot A, Uebele VN, Renger JJ, Eschalier A, Ardid D & M (2001). Zn2+ and H+ are coactivators of acid-sensing ion Bourinet E (2013). State-dependent properties of a new channels. J Biol Chem 276, 35361–35367. T-type calcium channel blocker enhance CaV3.2 selectivity Bernhardt ML, Zhang Y, Erxleben CF, Padilla-Banks E, and support analgesic effects. Pain 154, 283–293. McDonough CE, Miao YL, Armstrong DL & Williams CJ François A, Schuetter¨ N, Laffray S, Sanguesa J, Pizzoccaro A, (2015). CaV3.2 T-type channels mediate Ca2+ entry during Dubel S, Mantilleri A, Nargeot J, Noel¨ J, Wood John N, oocyte maturation and following fertilization. JCellSci128, Moqrich A, Pongs O & Bourinet E (2015). The low-threshold 4442–4452. calcium channel Cav3.2 determines low-threshold Blesneac I, Chemin J, Bidaud I, Huc-Brandt S, Vandermoere F mechanoreceptor function. Cell Reports 10, 370–382. & Lory P (2015). Phosphorylation of the Cav3.2 T-type Garcia-Caballero A, Gadotti VM, Stemkowski P, Weiss N, calcium channel directly regulates its gating properties. Proc Souza IA, Hodgkinson V, Bladen C, Chen L, Hamid J, Natl Acad Sci USA 112, 13705–13710. Pizzoccaro A, Deage M, Francois A, Bourinet E & Zamponi Bourinet E, Alloui A, Monteil A, Barrere C, Couette B, Poirot GW (2014). The deubiquitinating enzyme USP5 modulates O, Pages A, McRory J, Snutch TP, Eschalier A & Nargeot J neuropathic and inflammatory pain by enhancing Cav3.2 (2005). Silencing of the Cav3.2 T-type calcium channel gene channel activity. Neuron 83, 1144–1158. in sensory neurons demonstrates its major role in Grauert A, Engel D & Ruiz AJ (2014). Endogenous zinc nociception. EMBO J 24, 315–324. depresses GABAergic transmission via T-type Ca2+ channels Bourinet E, Altier C, Hildebrand ME, Trang T, Salter MW & and broadens the time window for integration of Zamponi GW (2014). Calcium-permeable ion channels in glutamatergic inputs in dentate granule cells. JPhysiol592, pain signaling. Physiol Rev 94, 81–140. 67–86. Boycott HE, Dallas ML, Elies J, Pettinger L, Boyle JP, Scragg JL, Hansen PB (2015). Functional importance of T-type Gamper N & Peers C (2013). Carbon monoxide inhibition of voltage-gated calcium channels in the cardiovascular and Cav3.2 T-type Ca2+ channels reveals tonic modulation by renal system: news from the world of knockout mice. Am J thioredoxin. FASEB J 27, 3395–3407. Physiol Regul Integr Comp Physiol 308, R227–237.

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society J Physiol 594.13 H191-dependent modulation of neuronal Cav3.2 channels 3573

Hilaire C, Lucas O, Valmier J & Scamps F (2012). Matsunami M, Kirishi S, Okui T & Kawabata A (2011). Neurotrophin-4 modulates the mechanotransducer Cav3.2 Chelating luminal zinc mimics hydrogen sulfide-evoked T-type calcium current in mice down-hair neurons. Biochem colonic pain in mice: possible involvement of T-type calcium J 441, 463–471. channels. Neuroscience 181, 257–264. Hirzel K, Muller¨ U, Latal AT, Hulsmann¨ S, Grudzinska J, Maturana A, Lenglet S, Python M, Kuroda S & Rossier MF Seeliger MW, Betz H & Laube B (2006). Hyperekplexia (2009). Role of the T-type calcium channel CaV3.2 phenotype of glycine receptor α1 subunit mutant in the chronotropic action of corticosteroids in mice identifies Zn2+ as an essential endogenous isolated rat ventricular myocytes. Endocrinology 150, modulator of glycinergic neurotransmission. Neuron 52, 3726–3734. 679–690. Nelson MT, Joksovic PM, Perez-Reyes E & Todorovic SM Hu H, Bandell M, Petrus MJ, Zhu MX & Patapoutian A (2009). (2005). The endogenous redox agent L-cysteine induces Zinc activates damage-sensing TRPA1 ion channels. Nat T-type Ca2+ channel-dependent sensitization of a novel Chem Biol 5, 183–190. subpopulation of rat peripheral nociceptors. JNeurosci25, Huang D, Huang S, Peers C, Du X, Zhang H & Gamper N 8766–8775. (2015). GABAB receptors inhibit low-voltage activated and Nelson MT, Joksovic PM, Su P, Kang HW, Van Deusen A, high-voltage activated Ca2+ channels in sensory neurons via Baumgart JP, David LS, Snutch TP, Barrett PQ, Lee JH, distinct mechanisms. Biochem Biophys Res Commun 465, Zorumski CF, Perez-Reyes E & Todorovic SM (2007a). 188–193. Molecular mechanisms of subtype-specific inhibition of Huc S, Monteil A, Bidaud I, Barbara G, Chemin J & Lory P neuronal T-type calcium channels by ascorbate. JNeurosci (2009). Regulation of T-type calcium channels: signalling 27, 12577–12583. pathways and functional implications. Biochim Biophys Acta Nelson MT, Woo J, Kang HW, Vitko I, Barrett PQ, Perez-Reyes 1793, 947–952. E, Lee JH, Shin HS & Todorovic SM (2007b). Reducing Jeong SW, Park BG, Park JY, Lee JW & Lee JH (2003). Divalent agents sensitize C-type nociceptors by relieving high-affinity metals differentially block cloned T-type calcium channels. zinc inhibition of T-type calcium channels. JNeurosci27, Neuroreport 14, 1537–1540. 8250–8260. Kang HW, Park JY, Jeong SW, Kim JA, Moon HJ, Perez-Reyes E Nevin ST, Cromer BA, Haddrill JL, Morton CJ, Parker MW & & Lee JH (2006). A molecular determinant of nickel Lynch JW (2003). Insights into the structural basis for zinc inhibition in Cav3.2 T-type calcium channels. J Biol Chem inhibition of the glycine receptor. J Biol Chem 278, 281, 4823–4830. 28985–28992. Kang HW, Vitko I, Lee SS, Perez-Reyes E & Lee JH (2010). Nozaki C, Vergnano AM, Filliol D, Ouagazzal AM, Le Goff A, Structural determinants of the high affinity extracellular zinc CarvalhoS,ReissD,Gaveriaux-RuffC,NeytonJ,PaolettiP binding site on Cav3.2 T-type calcium channels. J Biol Chem & Kieffer BL (2011). Zinc alleviates pain through 285, 3271–3281. high-affinity binding to the NMDA receptor NR2A subunit. Kawabata A, Ishiki T, Nagasawa K, Yoshida S, Maeda Y, Nat Neurosci 14, 1017–1022. Takahashi T, Sekiguchi F, Wada T, Ichida S & Nishikawa H Paoletti P, Ascher P & Neyton J (1997). High-affinity zinc (2007). Hydrogen sulfide as a novel nociceptive messenger. inhibition of NMDA NR1-NR2A receptors. JNeurosci17, Pain 132, 74–81. 5711–5725. Kuo CC, Chen WY & Yang YC (2004). Block of Perez-Reyes E (2003). Molecular physiology of tetrodotoxin-resistant Na+ channel pore by multivalent low-voltage-activated T-type calcium channels. Physiol Rev cations: gating modification and Na+ flow dependence. 83, 117–161. JGenPhysiol124, 27–42. Prost AL, Bloc A, Hussy N, Derand R & Vivaudou M (2004). Lee JH, Gomora JC, Cribbs LL & Perez-Reyes E (1999). Nickel Zinc is both an intracellular and extracellular block of three cloned T-type calcium channels: low regulator of KATP channel function. JPhysiol559, concentrations selectively block α1H. Biophys J 77, 157–167. 3034–3042. Reynders A, Mantilleri A, Malapert P, Rialle S, Nidelet S, Liao Y-F, Tsai M-L, Chen C-C & Yen C-T (2011). Involvement Laffray S, Beurrier C, Bourinet E & Moqrich A (2015). of the Cav3.2 T-type calcium channel in thalamic neuron Transcriptional profiling of cutaneous MRGPRD free nerve discharge patterns. Mol Pain 7, 43. endings and C-LTMRs. Cell Rep 10, 1007–1019. Liu T, Walker JS & Tracey DJ (1999). Zinc alleviates thermal Rose KE, Lunardi N, Boscolo A, Dong X, Erisir A, hyperalgesia due to partial nerve injury. Neuroreport 10, Jevtovic-Todorovic V & Todorovic SM (2013). 1619–1623. Immunohistological demonstration of CaV3.2 T-type Magistretti J, Castelli L, Taglietti V & Tanzi F (2003). Dual voltage-gated calcium channel expression in soma of dorsal effect of Zn2+ on multiple types of voltage-dependent Ca2+ root ganglion neurons and peripheral axons of rat and currents in rat palaeocortical neurons. Neuroscience 117, mouse. Neuroscience 250, 263–274. 249–264. Safieh-Garabedian B, Poole S, Allchorne A, Kanaan S, SaadEN´ Mathie A, Sutton GL, Clarke CE & Veale EL (2006). Zinc and & Woolf CJ (1996). Zinc reduces the hyperalgesia copper: pharmacological probes and endogenous and upregulation of NGF and IL-1β produced by modulators of neuronal excitability. Pharmacol Ther 111, peripheral inflammation in the rat. Neuropharmacology 35, 567–583. 599–603.

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society 3574 T. Voisin and others J Physiol 594.13

Sekiguchi F, Miyamoto Y, Kanaoka D, Ide H, Yoshida S, Weiss N, Hameed S, Fernandez-Fernandez JM, Fablet K, Ohkubo T & Kawabata A (2014). Endogenous and Karmazinova M, Poillot C, Proft J, Chen L, Bidaud I, exogenous hydrogen sulfide facilitates T-type calcium Monteil A, Huc-Brandt S, Lacinova L, Lory P, Zamponi GW channel currents in Cav3.2-expressing HEK293 cells. & De Waard M (2012). A Cav3.2/syntaxin-1A signaling Biochem Biophys Res Commun 445, 225–229. complex controls T-type channel activity and low-threshold Shcheglovitov A, Vitko I, Lazarenko RM, Orestes P, Todorovic exocytosis. J Biol Chem 287, 2810–2818. SM & Perez-Reyes E (2012). Molecular and biophysical basis White G, Lovinger DM & Weight FF (1989). Transient of glutamate and trace metal modulation of voltage-gated low-threshold Ca2+ current triggers burst firing through an Cav2.3 calcium channels. JGenPhysiol139, 219–234. afterdepolarizing potential in an adult mammalian neuron. Sheets MF & Hanck DA (1992). Mechanisms of extracellular Proc Natl Acad Sci USA 86, 6802–6806. divalent and trivalent cation block of the sodium current in Zamponi GW, Lory P & Perez-Reyes E (2010). Role of canine cardiac Purkinje cells. JPhysiol454, 299–320. voltage-gated calcium channels in epilepsy. Pflugers Arch Shin JB, Martinez-Salgado C, Heppenstall PA & Lewin GR 460, 395–403. (2003). A T-type calcium channel required for normal Zhang S, Kehl SJ & Fedida D (2001). Modulation of Kv1.5 function of a mammalian mechanoreceptor. Nat Neurosci 6, potassium channel gating by extracellular zinc. Biophys J 81, 724–730. 125–136. Talley EM, Cribbs LL, Lee JH, Daud A, Perez-Reyes E & Bayliss DA (1999). Differential distribution of three members of a gene family encoding low voltage-activated (T-type) calcium Additional information channels. JNeurosci19, 1895–1911. Todorovic SM & Jevtovic-Todorovic V (2007). Regulation of Competing interests T-type calcium channels in the peripheral pain pathway. The authors declare no conflict of interest. Channels (Austin) 1, 238–245. Todorovic SM & Jevtovic-Todorovic V (2013). Neuropathic pain: role for presynaptic T-type channels in nociceptive Author contributions signaling. Pflugers Arch 465, 921–927. Todorovic SM, Jevtovic-Todorovic V, Meyenburg A, Mennerick P.L.conceived the study. P.L.,E.B. and T.V.designed experiments. S, Perez-Reyes E, Romano C, Olney JW & Zorumski CF T.V.carried out all experimental work and data analysis. P.L.and (2001). Redox modulation of T-type calcium channels in rat T.V. wrote the paper. P.L.,E.B. and T.V. revised and approved the peripheral nociceptors. Neuron 31, 75–85. final version. Todorovic SM, Meyenburg A & Jevtovic-Todorovic V (2002). Mechanical and thermal antinociception in rats following systemic administration of mibefradil, a T-type calcium Funding channel blocker. Brain Res 951, 336–340. Traboulsie A, Chemin J, Chevalier M, Quignard JF, Nargeot J & T.V. was supported by a fellowship from the French Mini- Lory P (2006). Subunit-specific modulation of T-type stry of Research and Education. This work was supported by calcium channels by zinc. JPhysiol578, 159–171. Laboratory of Excellence ‘Ion Channel Science and Therapeutics’ Usoskin D, Furlan A, Islam S, Abdo H, Lonnerberg P, Lou D, (ANR-11-LABX-0015-01), by AFM grant (AFM-PainT) to E.B., Hjerling-Leffler J, Haeggstrom J, Kharchenko O, and by grants from Agence Nationale de la Recherche to Kharchenko PV, Linnarsson S & Ernfors P (2015). Unbiased E.B. (ANR-08-NMPS-025, ANR-09-MNPS-037) and to P.L. classification of sensory neuron types by large-scale (ANR-09-MNPS-035 and ANR-10-BLAN-1601). single-cell RNA sequencing. Nat Neurosci 18, 145–153. Vergnano AM, Rebola N, Savtchenko LP, Pinheiro PS, Casado M, Kieffer BL, Rusakov DA, Mulle C & Paoletti P (2014). Acknowledgements Zinc dynamics and action at excitatory synapses. Neuron 82, 1101–1114. We are grateful to Iulia Blesneac, Amaury François, Luc Wang R & Lewin GR (2011). The Cav3.2 T-type calcium Forichon, Sophie Laffray and Anne Pizzoccaro for technical channel regulates temporal coding in mouse support, as well as to Jean Chemin and Nathalie Guerineau´ for mechanoreceptors. JPhysiol589, 2229–2243. helpful comments.

C 2016 The Authors. The Journal of Physiology C 2016 The Physiological Society